RESEARCH ARTICLE

Stellate and pyramidal neurons in goldfish telencephalon responddifferently to anoxia and GABA receptor inhibitionNariman Hossein-Javaheri1, Michael P. Wilkie2, Wudu E. Lado3 and Leslie T. Buck1,*

ABSTRACTWith oxygen deprivation, the mammalian brain undergoes hyper-activity and neuronal death while this does not occur in the anoxia-tolerant goldfish (Carassius auratus). Anoxic survival of the goldfishmay rely on neuromodulatory mechanisms to suppress neuronalhyper-excitability. As γ-aminobutyric acid (GABA) is the majorinhibitory neurotransmitter in the brain, we decided to investigateits potential role in suppressing the electrical activity of goldfishtelencephalic neurons. Utilizing whole-cell patch-clamp recording, werecorded the electrical activities of both excitatory (pyramidal) andinhibitory (stellate) neurons. With anoxia, membrane potential (Vm)depolarized in both cell types from −72.2 mV to −57.7 mV and from−64.5 mV to −46.8 mV in pyramidal and stellate neurons,respectively. While pyramidal cells remained mostly quiescent,action potential frequency (APf) of the stellate neurons increased68-fold. Furthermore, the GABAA receptor reversal potential (EGABA)was determined using the gramicidin perforated-patch-clamp methodand found to be depolarizing in pyramidal (−53.8 mV) and stellateneurons (−42.1 mV). Although GABA was depolarizing, pyramidalneurons remained quiescent as EGABAwas below the action potentialthreshold (−36 mV pyramidal and −38 mV stellate neurons).Inhibition of GABAA receptors with gabazine reversed the anoxia-mediated response. While GABAB receptor inhibition alone did notaffect the anoxic response, co-antagonism of GABAA and GABAB

receptors (gabazine and CGP-55848) led to the generation ofseizure-like activities in both neuron types. We conclude thatwith anoxia, Vm depolarizes towards EGABA which increases APf instellate neurons and decreases APf in pyramidal neurons, andthat GABA plays an important role in the anoxia tolerance ofgoldfish brain.

KEY WORDS: Action potential, Spike frequency, Whole-cellconductance,Membranepotential,Actionpotential frequency,EGABA

INTRODUCTIONDuring periods of oxygen deprivation, ATP synthesis in the anoxia-intolerant brain (mammals) halts and compromises the ability ofthe Na+/K+-ATPase pump to maintain neuronal ion gradients,leading to membrane depolarization (Hansen, 1985). The release ofexcitatory neurotransmitters (glutamate, aspartate) follows, leadingto activation of α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic

acid (AMPA) and N-methyl-D-aspartate (NMDA) receptors (Sattlerand Tymianski, 2000). Subsequent calcium (Ca2+) influx throughNMDA receptors causes further depolarization of membranepotential (Vm), and activates intracellular Ca2+-dependentsignaling cascades, which leads to neuronal hyper-excitabilityand, eventually, excitotoxic cell death (Sattler and Tymianski,2000). As anoxia eliminates oxidative phosphorylation, theglycolytic pathway becomes the major source of ATP producedby the brain, generating less than 1/10th of the ATP producedunder normoxic conditions (Hochachka, 1986). The ATP shortfallis particularly problematic for neurons, as action potentialpropagation, consequent activation of output synapses andmaintenance of neuronal ion gradients are metabolicallyexpensive events, consuming approximately 60% of the totalbrain ATP turnover (Erecin ska and Silver, 1994). Anoxia-tolerantorganisms, such as the western painted turtle (Chrysemys picta),avoid anoxia-mediated cellular injury by not only reducing ATPsynthesis by over 90% (Bickler and Buck, 2007) but also reducingATP demand; indeed, in the case of turtle cortical pyramidalneurons, an anoxia-mediated 60–70% reduction in action potentialfrequency (APf) has been reported (Pamenter et al., 2011).

There are three hypotheses that encompass the events that lead toa large decrease in metabolic rate such as that exhibited by anoxic-tolerant species; these are the metabolic arrest (MA), channel arrest(CA) and spike arrest (SA) hypotheses (Hochachka, 1986; Lutz,1992; Buck and Hochachka, 1993). The MA hypothesis generallydescribes events leading to a reduction in ATP supply, while the CAhypothesis is specific to a reduction in ion channel conductanceleading to a reduction in ATP demand. The SA hypothesis isspecific to neurons or excitable tissues and entails a reduction in APfthat leads to a reduction in ATP demand and therefore a reduction inATP supply (Hochachka et al., 1996). Neurons function within anintegrated network; therefore, mechanisms regulating metabolic rateare also likely to occur at the network level. Specifically, with theonset of anoxia, some neurons may be activated in order to inhibitother neurons. We found evidence of this in anoxia-tolerant turtlecortical neurons, where glutamatergic pyramidal neurons undergospike arrest as a result of an increase in inhibitory γ-aminobutyricacid type A (GABAA) receptor currents (Pamenter et al., 2011).However, we have not recorded directly from the likely source of theincreased GABA levels – stellate interneurons.

Stellate neurons are smaller than pyramidal neurons and are moredifficult to patch clamp in the turtle cortex (Connors and Kriegstein,1986); therefore, our present goal was to develop a brain slicemodel inwhich stellate neurons could be reliably patched to obtain recordingsduring a normoxic to anoxic transition. As the reversal potential(EGABA) for the GABAA receptor is depolarized relative to Vm in theturtle pyramidal neurons, we also set out to determine EGABA in bothgoldfish stellate and pyramidal neurons. We hypothesized that, withthe onset of anoxia, stellate neuron Vmwill depolarize towardsEGABA,and APf will increase. Furthermore, we hypothesized that pyramidalReceived 22 July 2016; Accepted 30 November 2016

1Department of Cell and Systems Biology, University of Toronto, 25 Harbord St,Toronto, ON, Canada M5S 3G5. 2Department of Biology, Wilfred Laurier University,75 University Avenue West, Waterloo, ON, Canada N2L 3C5. 3Department ofNeurobiology, University of Alabama at Birmingham, 1825 University Blvd,Birmingham, AL 35294-2182, USA.

*Author for correspondence (les.buck@utoronto.ca)

L.T.B., 0000-0002-2637-7821

695

© 2017. Published by The Company of Biologists Ltd | Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

neuron Vm will depolarize towards EGABA, and whole-cellconductance will increase but APf will decrease.

MATERIALS AND METHODSAnimalsThis study was approved by the University of Toronto Animal CareCommittee and conforms to the relevant guidelines issued by theCanadian Council of Animal Care regarding the care and use ofexperimental animals. Common goldfish, Carassius auratus(Linnaeus 1758), weighing 50–100 g were purchased from acommercial supplier (AQUAlity Tropical Fish Wholesale,Mississauga, ON, Canada) and held in flowing dechlorinated Cityof Toronto tap water at 18°C at the University of Toronto.

Golgi stainingWhole goldfish brain was removed from the cranium followingdecapitation and the entire brain underwent Golgi staining. Thistechnique reveals fine structures of neurons, glia and their processes,allowing a better understanding of cellular morphology along withtheir orientation and location in cortical slices (Melendez-Ferroet al., 2009). Tissue was fixed in 10% formalin for 2 days followedby fixation in 3% K2Cr2O7 for 1 day and 2% AgNO3 for another2 days (Olude et al., 2015). The telencephalon was separated fromthe rest of the brain and sliced into 100 μm-thick axial slices using aLeica VT1200S vibratome (Leica, Richmond Hill, ON, Canada).Slices were collected using a soft brush, placed on a microscopeslide, cover-slipped and stored at 4°C.

Whole-cell electrophysiologyTelencephalon dissection, tissue preparation and whole-cell patch-clamp recordingmethods are described elsewhere (Wilkie et al., 2008).In short, thewhole brain of the goldfishwas removed from the craniumafter decapitation and was placed in oxygenated ice-cold 4°C artificialcerebrospinal fluid (aCSF) containing (in mmol l−1): 20 NaHCO3,118 NaCl, 1.2 MgCl2.6H2O, 1.2 KH2PO4, 1.9 KCl, 10 Hepes-Na,10 D-glucose and 2.4 CaCl2 (pH 7.6, adjusted with HCl); osmolarity285–290mOsm.Both lobes of telencephalonwere then dissected fromthe rest of the brain while in the chilled aCSF. Each telencephalon lobewas glued to a sectioning plate on a vibratome slicing chamber usingcyanoacrylate glue (Krazy Glue) and was filled with 4°C aCSF.Isolated goldfish telencephalon was cut into 300 µm slices using a

VT1200S vibratome. Slices were lifted out of the chamber and werestored in vials of oxygenated aCSF for no longer than 48 h. Slicesremained physiologically viable up to 48 h in the preparation (Wilkieet al., 2008). During each experiment, individual telencephalon sliceswere perfused with either oxygenated (99% O2, 1% CO2) or anoxic(99% N2, 1% CO2) aCSF (composition described above), andwhole-cell recordings were performed using the voltage-clampmethod with 9–11 MΩ borosilicate glass pipette electrodes (HarvardApparatus Ltd, Holliston, MA, USA) containing the following (inmmol l−1): 120 potassium gluconate, 10 KCl, 15 sucrose, 2 Na2ATP,0.44 CaCl2, 1 EGTA and 10 Hepes-Na (pH 7.6 adjusted with KOH);osmolarity 285–290 mOsm. The electrode was connected to a CV-4headstage (gain: 1/100 U, Axon Instruments, Sunnyvale, CA, USA)and an Axopatch 1D patch-clamp amplifier (Axon Instruments). Awhole-cell patch configuration was established by voltage clampingthe recordingpotential to−60 mVandapplying negative pressure aftera 1–20 GΩ seal was established. All seals were obtained using theblind-patch technique (Blanton et al., 1989). Whole-cell capacitancewas compensated, typical whole-cell access resistancewas 30–40MΩandseries resistancewasnot greater than50MΩ.Access resistancewasmonitored periodically and data were discarded if it changed by 10–15MΩ. Electrophysiological recordingswere performed onpyramidalneurons and stellate neurons in goldfish telencephalon. The cerebralcortex of fish, amphibians and reptiles is characterized by three layers,in which the pyramidal neurons are excitatory glutamatergic cells andthe stellate neurons are primarilyGABAergic and inhibitory, providingfeed-forward and feedback inhibition to pyramidal cells (Shepherd,2011; Larkum et al., 2008). Electrophysiologically, the two cell typeswere differentiated in goldfish telencephalon by their responses tosomatic current injections which produce trains of action potentials. Ifaccommodation occurred quickly, the cell was distinguished as apyramidal neuron and if no accommodationwas observed, the cell wasdifferentiated as a stellate neuron. Further, neural morphology wasdetermined by filling the microelectrode solution with 5% LuciferYellow in 2 mol l−1 LiCl. Lucifer Yellow was ejected by steadyhyperpolarizing current for 5 min after completion of the experiment(Connors and Kriegsten, 1986; Shin and Buck, 2003). Spontaneouselectrical activity was recorded for approximately 1 h (5–10 minnormoxia, 30 min anoxia, 20 min recovery) from pyramidal andstellate neurons. Action potential threshold (APth) was determined byinjecting current in a ramp manner in 1 nA steps until an AP waselicited. Whole-cell conductance (Gw) was measured by clampingneurons at voltage steps from −90 to −30 mV in 10 mV increments,lasting 200 ms each. Current amplitudes were measured between 135and 145 ms and a slope conductance according to the current–voltage(I/V ) relationship was determined (Pamenter et al., 2011).

Perforated-patch electrophysiologyWe utilized the gramicidin perforated-patch technique to determineEGABA without perturbing neural Cl− gradients. Micropipette tipswere front-filled with electrode recording solution by applyingsuction to the back of the pipette for 30 s, then back-filled with asolution of 13 µg ml−1 gramicidin in DMSO dissolved in electrodesolution. Gw recordings and I/V relationships were analyzed bytaking the linear regression of each line before and during2 mmol l−1 GABA application. EGABA was estimated bymeasuring the voltage at which the I/V relationships before andduring GABA application intersected (Watanabe et al., 2009).Using the Nernst equation and the known extracellular [Cl−], weestimated the intracellular [Cl−] from EGABA. EGABAwas calculatedby deducting conductance values pre- and post-GABA application,and plotting the results on a I/V curve.

List of abbreviationsaCSF artificial cerebrospinal fluidAMPA α-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acidAMPK AMP-dependent protein kinaseAPf action potential frequencyAPth action potential thresholdAPV (2R)-amino-5-phosphonovaleric acidBIC bicucullineCGP CGP-55848CNQX 6-cyano-7-nitroquinoxaline-2,3-dioneEGABA GABA reversal potentialEPSP excitatory post-synaptic potentialGABA γ-aminobutyric acidGIRK G-protein coupled inwardly rectifying potassium channelGw whole-cell conductanceGZ gabazine (SR95531)I/V current–voltageNMDA N-methyl-D-aspartateTTX tetrodotoxinVm membrane potential

696

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

PharmacologyGoldfish telencephalon slices were perfused with pharmacologicalmodifiers as specified in the Results through a fast-step perfusionsystem (VC-6 perfusion valve controller and SF-77B fast-stepperfusion system; Warner Instruments, Hamden, CT, USA).Possible involvement of the GABAergic system in the anoxicresponse was investigated by application of GABAA and GABAB

receptor antagonists, 25 μmol l−1 SR95531 gabazine (GZ),5 μmol l−1 CGP-55848 (CGP) and 100 μmol l−1 bicuculline(BIC) during anoxia. Changes in Vm and APf post-drug applicationwere compared. AMPA and NMDA receptors were blocked using25 μmol l−1 of 6-cyano-7-nitroquinoxaline-2,3-dione (CNQX) and25 μmol l−1 (2R)-amino-5-phosphonovaleric acid (APV),respectively. Voltage-gated sodium channels were blocked with1 μmol l−1 tetrodotoxin (TTX). EGABA was obtained by perfusingcortical slices with 2 mmol l−1 GABA. All chemicals were obtainedfrom Sigma-Aldrich Canada Ltd (Oakville, ON, Canada) exceptCGP and TTX, which were obtained from Tocris Cookson (Bristol,UK).

StatisticsElectrophysiological data were analyzed using SigmaPlot (version11.0; Systat Software, Inc., San Jose, CA, USA). All data weretested for normality and equal variance and an analysis of

variance (repeated measure ANOVA) was used for between-groupcomparisons where appropriate. Post hoc analyses consisted ofHolm–Sidak or Tukey’s test. Data are expressed as means±s.e.m.; nvalues represent a recording from a single cell from a singletelencephalic slice of a single goldfish. P<0.05 was consideredstatistically significant.

RESULTSGoldfish telencephalic neurons are divided into distinctmedially and laterally localized regionsGolgi staining and fluorescence imaging revealed that goldfishtelencephalon neurons were located in two distinct regions. The firstneural cell population was located at the lateral regions and cellswere spread into distinct layers (Fig. 1A). For simplicity, we dividedthe cells into three types. Type one cells had a triangular shapedsoma with dendritic fields extending away from the cell body inboth apical and basal directions (Fig. 1B,D). Their dendriticextensions were large with medial projections towards the center ofthe slice. With application of a suprathreshold current pulse, theseneurons showed rapid adaptation and fired distinct action potentialsfollowed by hyperpolarization (Fig. 1F). Cells with thesecharacteristics were identified in painted turtle cortex as pyramidalneurons (Connors and Kriegsten, 1986). Type two cells had a small(∼25 µm) and circular soma with bi-lateral projections extending

–83 mV

20 mV50 ms

–70 mV

A

Eii

F

G

D

EiB

C

Fig. 1. Goldfish telencephalic neurons. (A) An example of a goldfish telencephalon slice following Golgi staining, enlarged 10×. Cells were concentrated at twodistinct medial and lateral regions. The two populations were clearly separated by white matter and connective tissue. Area dorsalis telencephali pars centralis,indicated by the black arrow, was the medial region of the slice, which contained a large population of pyramidal neurons. The lateral regions, indicated bywhite arrow, contained both small, localized and large stellate cells withmedial projections. (B) A pyramidal cell magnified 40×. These cells showed two apical andbasal outward projections from their triangular shaped soma. (C) A group of stellate cells with circular shaped somaswere largely concentrated closer to the edgesand appeared to be divided into layers. (D) A pyramidal neuron filled with Lucifer Yellowand imaged immediately after whole-cell recording. These cells expressedthe same morphology observed with Golgi staining. (Ei) A stellate cell with a large dendritic tree and bi-lateral projections. (Eii) Small stellate neurons thatwere more concentrated at the lateral region. (F,G) Action potential properties of pyramidal (F) and stellate (G) neurons recorded under current clamp. Thepyramidal cells are rapidly adapting neurons while stellate neurons are slowly adapting and can generate action potentials rapidly in response to injected current.The scale bars represent 50 µm.

697

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

away from the cell body (Fig. 1C). It appeared that these cells weremore localized and had little branching on their dendritic tree(Fig. 1Eii). Type three cells had a more oval-shaped cell body,which was larger than type two cells (∼50 µm) (Fig. 1Ei). Similar topyramidal neurons, they sent their projections medially and hadlarge dendritic fields. Physiologically, both type two and type threecells showed little or no spike frequency accommodation inresponse to suprathreshold current. Action potentials weregenerated continuously and had equal amplitude (Fig. 1G). Thisphysiological morphology has been associated with stellate neurons(Connors and Kriegstein, 1986).The second population was part of the area dorsalis telencephali

pars centralis located at the center of the slice and was isolated fromthe lateral region (Peter and Gill, 1975). The area dorsalistelencephali pars centralis contained both pyramidal and type threestellate neurons; however, no type two stellate neurons wereobserved. Perhaps because of their small size, type two stellateneurons were encountered only twice. Hence, data collected fromthese cells were excluded from the data set. Both pyramidal andstellate cells had large dendritic trees and were highly interconnected.In this experiment, the term ‘stellate neurons’ refers to the type threestellate cells.

Pyramidal and stellate neurons respond differently to anoxiaPyramidal neurons are the primary excitatory neurons in the turtlecortex and are the major consumers of ATP (Howarth et al., 2010).A reduction in excitability of these neurons is crucial to tolerance ofanoxia and ischemia, and energy conservation (Pamenter et al.,2011; Hogg et al., 2015). Inhibition is not independent from the restof the cortex and if GABAergic stimulation is responsible forsuppression of neuronal excitability, the inhibitory stellate cellsshould show greater excitation with anoxia. Under passiverecording conditions (I=0), it was observed that pyramidalneurons were mostly quiescent and Vm was maintained at −72.18±2.3 mV during normoxia. Vm depolarized to −57.7±1.6 mV after30 min of anoxia (P=0.01, n=14) and repolarized to −64.22±2.6 mV after 20 min of normoxic recovery (P=0.08, n=14). DespiteVm depolarization, these cells did not generate an action potential(Fig. 2A,E,F). However, a small population of pyramidal neuronswere more depolarized and active during normoxia (Vm=−60.5±1.6 mV, APf=0.51±0.03 Hz). In these cells, anoxic treatmentcaused a depolarization of Vm to −53±1.08 mV (P=0.04, n=4)and APf decreased to 0.1±0.06 Hz (P<0.001, n=4). With normoxicrecovery, Vm hyperpolarized to −55±2.6 mV; however, this was notsignificantly different from anoxic Vm; APf increased to 0.24±0.1 Hz (P=0.003 from anoxia to recovery, n=4) (Fig. 2D,G,H).Similar to findings in pyramidal neurons, Vm depolarized from−64.54±1.05 mV to −46.86±1.2 mV and repolarized to −53±2.65 mV in stellate neurons during a normoxic to anoxic transitionand recovery (P=0.02 for normoxia to anoxia and P=0.03 for anoxiato recovery, n=11). Unlike pyramidal neurons, stellate neuronsgenerated action potentials with anoxia and APf increased from0.03±0.009 Hz to 2.04±0.12 Hz (P<0.001, n=11). APf was reducedto 0.13±0.03 Hz with normoxic recovery (P<0.001, n=11) (Fig. 2B,E,F). Application of AMPA and NMDA receptor antagonists(25 μmol l−1 CNQX and 25 μmol l−1 APV, respectively) did notsuppress action potentials; however, 1 μmol l−1 TTX abolishedthem (data not shown). Furthermore, both pyramidal and stellateneurons showed an ∼60% increase in Gw with anoxia. Thesechanges were from 1.75±0.34 nS to 2.83±0.17 nS in pyramidalneurons (P=0.04, n=7) and 1.27±0.15 nS to 2.01±0.16 nS in stellatecells (P=0.07, n=8) (Fig. 2G).

Vm shifts towards EGABA with anoxiaActivation of GABAA receptors leads to a Cl− flux in the directionof the reversal potential (ECl) (Blaesse et al., 2009). As GABAA

receptors are predominantly permeable to Cl−, EGABA and ECl

should be similar (Errington et al., 2014). We measured EGABA ingoldfish telencephalic neurons to better understand the relationshipbetween Vm and EGABA. In order not to perturb neuronal Cl−

gradients, we used the gramicidin perforated-patch technique todetermine EGABA. Addition of 2 mmol l−1 GABA during normoxiaincreased Vm in telencephalic neurons by 20±2.6 mV (n=8 forpyramidal neurons and n=8 for stellate neurons, P<0.001), whichreturned to baseline approximately 10 min post-GABA application(Fig. 3A). In pyramidal neurons, stepwise current injectiongenerated action potentials in the absence of GABA but not in thepresence of GABA (Fig. 3B). Similarly, stellate neurons showedsuppressed activity with GABA application; however, withincreased current injection, these neurons were capable ofgenerating action potentials (Fig. 3C). Furthermore, Gw wasrecorded before and after GABA application and a I/Vrelationship was constructed. GABA increased Gw and theintersection between the two I/V curves was taken as EGABA

(Fig. 3D). EGABA was more depolarized in stellate neurons than inpyramidal cells: −42.1±2.3 and −53.80±2.63 mV, respectively. Byobtaining the EGABA values and utilizing the Nernst equation, wewere able to calculate intracellular [Cl−], which was ∼15.1 and∼24 mmol l−1 in pyramidal and stellate neurons, respectively. Asdemonstrated previously, Gw increased with anoxia from 3.96±0.95 nS to 6.84±0.61 nS (P<0.001, n=5). Addition of GABAappeared to increase Gw during normoxia; however, this increasewas not significant (Gw=5.49±0.86 nS, P=0.3, n=5). With anoxia,Gw increased to 8.49±0.2 nS, which was significantly differentfrom anoxia without GABA (P=0.02, n=5) and from normoxia withGABA (P<0.001, n=5) (Fig. 3E).

Anoxic Vm depolarization is reversed by inhibition of GABAAand GABAB receptorsIn order to investigate the possible role of GABA in the defenseagainst anoxia-mediated damage in goldfish telencephalic neurons,we blocked GABAA and GABAB receptors under anoxicconditions. During anoxia, pyramidal neuron Vm hyperpolarizedwith application of the GABAA receptor inhibitor GZ (25 μmol l−1)from −56.54±1.7 mV to −64.77±1.9 mV and depolarized back to−55.15±1.36 mVwith normoxic recovery (P<0.001, n=6) (Fig. 4A,D).In anoxic stellate neurons, GZ also hyperpolarized Vm, from−46.8±2.1 mV to −58.54±2.1 mV (P<0.001, n=5) (Fig. 4B,D)and APf decreased from 2.04±0.12 Hz to 0.155±0.01 Hz (P<0.001,n=5) (Fig. 4E). These changes were reversed with the terminationof GZ application. The GABAA receptor inhibitor BICalso hyperpolarized Vm to −68.82±1.7 mV and −56.8±1.97 mV in pyramidal and stellate neurons, respectively (n=3,P<0.001; data not shown). No significant difference was observedbetweenBIC andGZ in reversal of the anoxic response. Even thoughtheGABAB receptor antagonist CGPalone did not changeVm orAPfin either cell type (Vm=−53.97±2.24 mV for pyramidal and Vm=−47.74±2.83 mV, n=4), co-antagonism of GABAA and GABAB

receptors with GZ plus CGP significantly influenced neuronalexcitability (Fig. 4E). GZ+CGP initially hyperpolarized Vm similarto what was observed with GZ perfusion. However, shortly after, Vm

depolarized significantly (Vm=−36.58±1.62, P<0.001, n=4, and−34.25±5.9 mV, P<0.001, n=5 for pyramidal and stellate neurons,respectively) and generated sudden, seizure-like action potentials(Fig. 4C,D). Furthermore, APf increased in pyramidal neurons from

698

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

Normoxia(5 min)

Anoxia(30 min)

Recovery(20 min)

–75 mV

–65 mV

Normoxia (10 min)(Iinj=1 nA) (Iinj=1 nA)

(Iinj=1 nA) (Iinj=1 nA)Anoxia (30 min)

–65 mV–75 mV

–60 mV –50 mV20 mV

50 ms

Normoxia (10 min)

Anoxia (30 min)

A

B

Ci Cii Normoxia (1 min)

Anoxia (1 min)

20 m

V

10 s

D

Normoxia(5 min)

Anoxia(30 min)

Recovery(20 min)

10 m

V

5 min

–80

–70

–60

–50

–40

–30

–20

–10

0

Vm

(mV

)A

Pf (

Hz)

Pyramidal Stellate

Pyramidal Stellate

Normoxia RecoveryAnoxia

Pyramidal

Stellate

0

0.1

0.2

0.3

0.4

0.5

0.6

AP

f (H

z)

–70

–60

–50

–40

–30

–20

–10

0

Vm

(mV

)

E

F H

0

0.5

1

1.5

2

2.5

3

3.5

Gw (n

S)

0

0.5

1

1.5

2

2.5

G

**

*

‡

*** ‡‡‡

*

‡‡**

Normox

ia

Anoxia

Recov

ery

Normox

ia

Anoxia

Recov

ery

Normox

ia

Anoxia

Recov

ery

Fig. 2. See next page for legend.

699

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

quiescent to 2.23±0.09 Hz (n=4, P<0.001); the increase in APf wasnot significantly different in stellate neurons (APf=2.13±0.02 Hz,n=5) (Fig. 4E). Even though APf drastically increased with GZ+CGP application, in some neurons (n=5) Vm hyperpolarized tothe pre-anoxia level, neurons became quiescent and stepwisecurrent injection failed to generate an action potential (data notshown). The rest (n=4), however, did not recover afterGZ+CGP application and continued to generate seizure-like actionpotentials until loss of Vm, which is associated with neuronal death(Fig. 4C).GABAA/B receptor antagonism significantly affected Gw in both

pyramidal and stellate neurons. Anoxia increased Gw from 3.96±0.95 nS to 6.84±0.62 nS (P<0.001). Blocking the GABAA receptorwith GZ had no significant effect on normoxic Gw (2.83±0.33 nS,P=0.071) but it did significantly block the anoxia-mediated increase(3.87±0.39 nS, P<0.001). Normoxic application of CGP alone didnot have an effect on Gw (4.1±0.56 nS, P=0.81) and application ofCGP alone during the anoxic period did not suppress the anoxia-mediated increase (5.43±0.59 nS, P=0.27). Co-application of GZ+CGP did not affect normoxic Gw (3.48±0.18 nS, P=0.43) butdid significantly block the anoxia-mediated increase in Gw (3.7±0.26 nS, P<0.001). This suggests that the increase in conductanceduring anoxia is dependent on the activation of GABAA receptorsand that GABA is likely responsible for anoxic depolarization(Fig. 4F). Gw measurements were conducted on anoxic andnormoxic controls and following treatment with GZ, CGP and GZ+CGP, each with 5 replicates. As the change inGw was similar for thetwo neuronal types, the pharmacological responses were combined.Data are presented as the combined responses of pyramidal andstellate neurons.

DISCUSSIONIn this study we investigated the role of GABAergicneurotransmission in the telencephalon of the naturally anoxia-tolerant goldfish,C. auratus. We show that anoxia depolarizes the Vm

of both stellate and pyramidal neurons towards the GABAA receptorreversal potential (EGABA), which is reversed by application of aGABAA receptor blocker. Interestingly, depolarization by increasingGABAA receptor currents decreases pyramidal cell APf while itincreases the APf of stellate neurons. Furthermore, we show GABAB

receptors are also important in anoxic survival as inhibition of bothGABAA and GABAB receptors induce seizure-like activities intelencephalic neurons and apparent cell death. GABA is the primaryinhibitory neurotransmitter in the mature mammalian central nervoussystem and an increase in the extracellular [GABA] seems to be acommon response to anoxia in anoxia-tolerant organisms (Krnjevic,1997; Nilsson and Lutz, 2004; Bickler and Buck, 2007). In theanoxia-tolerant turtle brain (Chrysemys picta bellii, Trachemysscripta elegans), [GABA] increases by 80-fold, suppressing whole-brain electrical activity by 75–95% (Lutz and Nilsson, 1997). This isstrongly mediated by both GABAA and GABAB receptors, which arecritical to anoxic survival of turtle cortical neurons (Pamenter et al.,2011). While turtles become quiescent and enter a stage of minimalactivity, goldfish and crucian carp remain slightly active in anoxicwaters (Bickler and Buck, 2007). Compared with levels in turtlebrain, the anoxia-mediated increase in extracellular [GABA] incrucian carp brain is quite modest and doubles after 5 h of anoxia(Hylland and Nilsson, 1999). Furthermore, with anoxia, GABAremoval from the extracellular space is significantly reduced becausethe expression of GABA transporter proteins decreases by 80%,increasing the effective extracellular [GABA] further (Ellefsen et al.,2009). We observed a significant increase in neuronal Vm withanoxia. This depolarization led to action potential propagation instellate neuronswhile pyramidal neurons remained silent. An increasein AP frequency of stellate neurons seems counter to energyconservation within a low oxygen environment; however, inhibitoryneurons consume significantly less energy than excitatory neuronsand there are about 10-fold fewer (Howarth et al., 2010). In a study byHylland and Nilsson (1999), running a high [K+] Ringer solutionthrough a microdialysis probe to depolarize the surrounding tissuecaused the extracellular [GABA] to increase 14 times, whileextracellular [glutamate] doubled. This suggests that the potentialfor GABA release in crucian carp telencephalon is significantlygreater than that for glutamate, resulting in greater GABA release inresponse to equal amounts of stimulation. If we consider thesimilarities between the telencephalon of teleost fish and mousehippocampus, the number of excitatory synapses is approximately18-fold greater than the number of inhibitory ones (Megías et al.,2001; Muller, 2012). Therefore, as a defense mechanism in goldfishtelencephalic neurons, a small population of neurons with highrelease potential and lower energy requirements are activated tosuppress the larger population of excitatory neurons and conserveenergy.

In our experiment, blocking NMDA and AMPA receptors failedto suppress action potential firing in stellate neurons while TTXblocked action potentials completely. Therefore, we hypothesizedthat pathways other than excitatory glutamatergic signaling, such asGABAergic pathways, are responsible for Vm depolarization duringanoxia. Indeed, blocking GABAA receptors with GZ reversedanoxic depolarization of Vm and prevented the anoxia-mediatedincrease in Gw. Under physiological conditions, GABAA receptorsare highly permeable to both Cl− and HCO3

−; however, thechannels are considerably more permeable to Cl− than to HCO3

−

(Cl−/HCO3− permeability: 5/1) and, thus, the majority of charge

transfer following channel activation is Cl− mediated (Erringtonet al., 2014). Therefore, we conclude that the anoxic increase in Gw

is largely due to the movement of anions.Measurements of EGABA revealed that Vm shifts towards a

depolarizing EGABA with anoxia and induces action potentials instellate neurons but not in pyramidal cells. This is consistent with aprevious report showing exogenous GABA can depolarize goldfishtelencephalic gonadotropin-releasing hormone (GnRH) neurons

Fig. 2. Basic electrophysiological properties of pyramidal and stellateneurons. (A) Sample trace of a free current-clamp recording from a pyramidalneuron, demonstrating changes in the membrane potential (Vm) duringnormoxia, anoxia and recovery. (B) Sample trace of a free current-clamprecording from a stellate neuron showing changes in Vm and action potentialfrequency (APf). (C) Sample recordings of action potentials by stepwisecurrent injection. (i) In pyramidal neurons, anoxia decreases the probability ofaction potential generation in response to the same amount of stimulation(1 nA). (ii) In stellate neurons, the probability of action potential firing does notchange with anoxia; as a result of Vm depolarization, firing may increase.(D) Occasionally, an active pyramidal neuron was found (n=4); example 1 mintraces of spontaneous firing of the active neuron during normoxia and then5 min of anoxia are shown. (E–G) Summary of changes in: Vm (E), APf (F) andwhole-cell conductance (Gw; G) in both pyramidal and stellate neurons fornormoxia, anoxia and recovery. With anoxia, Gw increased in both cell types;Gw did not change significantly with recovery in either cell type. (H) Changesin Vm and APf of active pyramidal neurons as in D. Treatments: normoxia:99%O2, 1%CO2 bubbled artificial cerebrospinal fluid (aCSF); anoxia: 99%N2,1% CO2 bubbled aCSF; and recovery: 99% O2, 1% CO2 bubbled aCSF. Dataare means±s.e.m., n=7–14 replicates per treatment and n=4 for activepyramidal neurons. Asterisks indicate a significant difference from baselinenormoxia; double daggers indicate a significant difference from normoxicrecovery in pyramidal and stellate neural populations: */‡P<0.05, **/‡‡P<0.01,***/‡‡‡P<0.001. n.s., not significant.

700

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

(Nakane and Oka, 2010). By changing Vm to EGABA in pyramidalneurons, the potential difference between the intracellular andextracellular environments is decreased, which significantly reducesenergy in the form of ATP required to maintain ionic gradients(Edwards et al., 1989). Although GABA is depolarizing, it remainsinhibitory because it activates an inhibitory shunt that maintains Vm

below the threshold potential (Pamenter et al., 2011). EGABA isprimarily determined by the Cl− gradient where K+/Cl−-cotransporter 2 (KCC2) and Na+/K+/Cl−-cotransporter 1(NKCC1) set the reversal potential and driving force of anioncurrents mediated by GABAA receptors (Blaesse et al., 2009). ForGABA to have a depolarizing effect, the intracellular [Cl−] must berelatively high (20–25 mmol l−1) so that an efflux of Cl− occurs viaGABAA receptors (Ben-Ari, 2002). We found that intracellular[Cl−] is relatively high in goldfish telencephalic neurons andapproximately 1.5 times greater in stellate neurons than in pyramidalneurons. As a result, activation of GABAA receptors leads to agreater efflux of Cl− and further depolarization in stellate neuronscompared with pyramidal neurons. Cl− flow occurs through both

synaptic (phasic) and extrasynaptic (tonic) GABA receptors(Blaesse et al., 2009). Localization of GABAA receptors has notbeen investigated in goldfish neurons; however, crucian carpneurons express the gamma subunit, which is necessary for phasiccurrents, and delta subunits, which in mammalian brain areextrasynaptic (Mody, 2001; Ellefsen et al., 2009). By applyingGZ, we likely inhibited the activity of phasic currents as GZ has agreater affinity for synaptic GABAA receptors (Mody, 2001).Application of BIC hyperpolarized Vm of both pyramidal andstellate neurons and reduced APf in stellate neurons similar to GZ.However, BIC blocks both phasic and tonic GABAA receptors(Ueno et al., 1997), suggesting that the anoxic response is primarilymediated through the activity of the synaptic receptors. Anoxia-tolerant turtle brain differs from goldfish brain as anoxia induceslarge phasic GABA currents and a 50% increase in tonic current(Hogg et al., 2015). Large phasic GABAergic currents were onlyobserved in a single goldfish pyramidal neuron recordings (1/47attempts). Perhaps formation of such currents requires an intactcircuitry or is localized to a specific cortical region. As our

A

B

D

GABA

–80 mV

–72 mV –63 mV

20 m

V

50 ms

–67 mV

+GABA +GABA

–75 mV –67 mV10 mV

30 s

Pyramidal neuron Stellate neuron

0

1

2

3

4

5

6

7

8

9

10

Gw (n

S)

Normoxia

Anoxia

Normoxia+GABA

Anoxia+GABA

*******E

C

–200

–150

–100

–50

0

50

100

–100 –90 –80 –70 –60 –50 –40 –30 –20 –10 0

I (pA

)

V (mV)

Pyramidal neurons

Stellate neurons

GABA

Fig. 3. Electrophysiological responses of pyramidal and stellate neurons toGABA application. (A) Passive recordings from pyramidal and stellate neuronsduring normoxia in the absence and presence of GABA. In both neuronal types, application of GABA increases Vm during normoxia and Vm recovers to theinitial baseline level minutes after the termination of GABA application. (B,C) Sample recordings of action potentials from pyramidal neurons (B) and stellateneurons (C) following stepwise current injection before and after GABA application during normoxia. (D) Whole-cell current–voltage (I/V ) relationship frompyramidal and stellate neurons in response to GABA. Application of GABA shifts Vm towards the GABA reversal potential (EGABA). (E) Changes in Gw in bothpyramidal and stellate neurons with and without exogenous GABA application during normoxia and anoxia. Treatments: normoxia: 99% O2, 1% CO2 bubbledaCSF; [GABA], 2 mmol l−1. Data are means±s.e.m., n=5–8 replicates per treatment. Asterisks indicate significance: *P<0.05, **P<0.01 and ***P<0.001.

701

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

A

B

Ci CiiPyramidal neuron

Normoxia–74 mV

Anoxia RecoveryGZ

Normoxia Anoxia Recovery

20 m

V4 min

GZ

–57 mV

Anoxia Anoxia+GZ+CGP

–47 mV

Anoxia Anoxia+GZ+CGP

ab b

–47 mV

a

40 m

V

–28 mV

b

15 m

V

2 minStellate neuron

–80

–70

–60

–50

–40

–30

–20

–10

0Ano

xia

GZ CGPCGP+G

Z

Anoxia

GZ CGPGZ+C

PG

Contro

l

GZ CGPGZ+C

PG

Vm

(mV

)

Pyramidal Stellate

0

0.5

1

1.5

2

2.5

AP

f (H

z)

Pyramidal Stellate

0

1

2

3

4

5

6

7

8

Gw (n

S)

Normoxia

Anoxia

******

****** ***

***

D E

F

*** ******

n.s.

–66 mV

20 m

V

Fig. 4. See next page for legend.

702

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

electrophysiological recordings were from brain slices, the circuitrymight have been disrupted (Dukoff et al., 2014; Hogg et al., 2015).Furthermore, all the recordings utilized the blind-patch technique,which prevents clear localization of the areas populated withneurons that induce large GABA-mediated currents (Blanton et al.,1989). As a result of such limitations and the inability to frequentlyrecord GABA-induced currents, it is not possible to state to whatdegree tonic GABAA receptors are involved in the anoxic responsein the goldfish brain.Blocking GABAB receptors alone did not affect Vm or APf and

changes in Gw were not as strong as those during GABAA receptorinhibition. These receptors are slow acting (100–500 ms) G-proteincoupled inwardly rectifying potassium channels (GIRKs) andmediate prolonged neural inhibition by increasing K+ efflux via K+

channels post-synaptically and inhibiting voltage-gated Ca2+

channels pre-synaptically (Gassmann and Bettler, 2012). GABAB

receptor activity is strongly mediated by intracellular messengermechanisms (Kuramoto et al., 2007). With anoxia or a shortage ofATP, certain intracellular signaling proteins, such as AMP-dependent protein kinase (AMPK), are activated due to anincrease in [AMP] (Carling, 2004). AMP has a critical role inmaintaining cellular and systemic energy balance and promotingglycolysis (Stensløkken et al., 2008). Through phosphorylation,AMPK strongly enhances GABAB receptor activity, allowing agreater K+ efflux via GIRKs (Kuramoto et al., 2007). Anoxicexposure significantly increases AMPK phosphorylation in cruciancarp brain (Stensløkken et al., 2008).In this study, we were able to show that inhibition of both

GABAA and GABAB receptors leads to sudden, uncontrollableelectrical activity and a paroxysmal depolarizing shift. Theparoxysmal depolarizing shift is a manifestation of neuralepilepsy that is in contrast to endogenous burst firing as it isnetwork driven (Johnston and Brown, 1984). In some neurons, Vm

did not return to baseline and seizure-like activities persisted. Inothers, Vm hyperpolarized to a pre-anoxic stage after hyperactivity

and cells became quiescent. This does not necessarily imply thatthese neurons recovered from seizure as this can occur as a result of adepolarization block in which seizure can be maintained even in theabsence of continuous neuronal firing (Bikson et al., 2003). Whileco-antagonism of GABAA and GABAB receptors inducedparoxysmal depolarizing shift, blocking either GABAA orGABAB receptors alone did not cause catastrophic depolarization.GABAA and GABAB receptors have a complementary role inreducing neural excitation (Mann et al., 2009). GZ significantlyreduced conductance and reversed the anoxic depolarization;however, GABAB receptors might prevent the release ofexcitatory neurotransmitters and excitotoxic cell death even whenGABAA receptors are inactive. In the anoxic turtle brain, GABAB

receptors predominantly act pre-synaptically and decreaseexcitatory post-synaptic potentials (EPSPs), especiallyAMPAergic EPSPs (Pamenter et al., 2011). Furthermore, weshow that inhibiting GABAB receptors alone does not significantlychange the anoxic Gw while blocking GABAA receptors does. Thissuggests that the major change in conductance brought on by anoxiamust occur through GABAA receptors and could explain whyapplication of CGP alone causes no catastrophic depolarization.Activity of GABAB receptors may be crucial to neural survivaland limiting excitotoxicity, especially if GABAA receptorsare inactive, perhaps by decreasing EPSPs and reducingneurotransmitter release (Kuramoto et al., 2007; Pamenter et al.,2011; Tao et al., 2013).

In conclusion, we report that in anoxic goldfish telencephalicstellate neurons, APf increases while pyramidal neuron APf or thepotential to generate action potentials is decreased. In both neuronaltypes, Vm depolarizes towards the EGABA but because EGABA ismore depolarizing in stellate neurons, it is responsible for theincreased activity. Inhibiting both GABAA and GABAB receptorsinduces terminal depolarization of goldfish neurons, highlightingthe importance of GABA in anoxia tolerance.

AcknowledgementsThe authors would like to thank Dr Kaori Takehara-Nishiuchi and Dr MartinWojtowicz for their valuable insights and critiques during every stage of this study.

Competing interestsThe authors declare no competing or financial interests.

Author contributionsConceptualization: L.T.B, N.H-J.; Methodology: N.H-J., W.E.L.; Formal analysis andinvestigation: N.H-J., W.E.L., L.T.B.: Writing - original draft preparation: N.H-J.:Writing - review and editing: L.T.B., N.H-J.; Funding acquisition: L.T.B., M.P.W.;Resources: L.T.B.; Supervision: L.T.B.

FundingThis study was supported by the Natural Sciences and Engineering ResearchCouncil (NSERC Discovery Grant) to L.T.B. (06588) and M.P.W. (04248).

ReferencesBen-Ari, Y. (2002). Excitatory actions of GABA during development: the nature of

the nurture. Nat. Rev. Neurosci. 3, 728-739.Bickler, P. E. and Buck, L. T. (2007). Hypoxia tolerance in reptiles, amphibians and

fishes: life with variable oxygen availability. Annu. Rev. Physiol. 69, 145-170.Bikson, M., Hahn, P. J., Fox, J. E. and Jefferys, J. G. R. (2003). Depolarization

block of neurons during maintenance of electrographic seizures. J. Neurophysiol.90, 2402-2408.

Blaesse, P., Airaksinen, M. S., Rivera, C. and Kaila, K. (2009). Cation-chloridecotransporters and neuronal function. Neuron 61, 820-838.

Blanton, M. G., Lo Turco, J. J. and Kriegstein, A. R. (1989). Whole cell recordingfrom neurons in slices of reptilian and mammalian cerebral cortex. J. Neurosci.Methods 30, 203-210.

Buck, L. T. and Hochachka, P. W. (1993). Anoxic suppression of Na(+)-K(+)-ATPase and constant membrane potential in hepatocytes: support forchannel arrest. Am. J. Physiol. 265, 1020-1025.

Fig. 4. Changes in electrophysiological properties in response to GABAreceptor inhibition during anoxia. (A) In pyramidal neurons, application ofthe GABAA receptor inhibitor gabazine (GZ) re-polarized Vm back to normoxiclevels during anoxic exposure. (B) In stellate neurons, similar to pyramidalneurons, GZ re-polarized Vm and reduced APf. Vm and APf returned to pre-GZstate with normoxic recovery. (C) Although application of the GABAB receptorinhibitor CGP-55845 (CGP) alone did not significantly affect Vm or APf (datanot shown), the combination of GZ+CGP initially hyperpolarized Vm in bothpyramidal (i) and stellate (ii) neurons, similar to that observed with GZapplication. With continuous GZ+CGP application, Vm depolarized in both celltypes and led to the generation of sudden seizure-like action potentials.Enlarged traces are shown below; however, only representative traces fromstellate neurons are shown because pyramidal neurons were quiescent beforedrug application and the response was similar after. (D) Changes in Vm instellate and pyramidal neurons during anoxia, with application of GZ, CGP andGZ+CGP. GZ hyperpolarized Vm towards pre-anoxia levels while CGP alonehad no significant effect on Vm. GZ+CGP significantly depolarized Vm. (E) APf

changes with GABA receptor inhibition. GZ had no significant effect on APf inpyramidal neurons; however, it significantly reduced APf in stellate cells. CGPhad no significant effect on APf changes in either cell type. GZ+CGP inducedseizure-like activities in both neuronal types, significantly increasing the APf inpyramidal cells. (F) Changes inGw in both stellate and pyramidal neurons.Withanoxia, Gw increased significantly. GZ and GZ+CGP significantly reduced theanoxic Gw while CGP alone had no significant effect on anoxic conductance.No significant change was observed following drug application duringnormoxia (Gw data were recorded immediately after drug application).Treatments: normoxia: 99% O2, 1% CO2 bubbled aCSF; anoxia: 99% N2,1% CO2 bubbled aCSF. [GZ], 25 μmol l−1, [CGP], 5 μmol l−1. n=5–9 for eachtreatment. Asterisks indicate significance: *P<0.05, **P<0.01 and ***P<0.001.n.s., not significant.

703

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology

Carling, D. (2004). The AMP-activated protein kinase cascade–a unifying systemfor energy control. Trends. Biochem. Sci. 29, 18-24.

Connors, B.W. and Kriegstein, A. R. (1986). Cellular physiology of the turtle visualcortex: distinctive properties of pyramidal and stellate neurons. J. Neurosci. 6,164-177.

Dukoff, D. J., Hogg, D. W., Hawrysh, P. J. and Buck, L. T. (2014). ScavengingROS dramatically increases NMDA receptor whole cell currents in painted turtlecortical neurons. J. Exp. Biol. 217, 3346-3355.

Edwards, R. A., Lutz, P. L. and Baden, D. G. (1989). Relationship between energyexpenditure and ion channel density in the turtle and rat brain. Am. J. Physiol. 257,1354-1358.

Ellefsen, S., Stenslokken, K.-O., Fagernes, C. E., Kristensen, T. A. and Nilsson,G. E. (2009). Expression of genes involved in GABAergic neurotransmission inanoxic crucian carp brain (Carassius carassius). Physiol. Genomics. 36, 61-68.

Erecinska, M. and Silver, I. A. (1994). Ions and energy in mammalian brain. Prog.Neurobiol. 43, 37-71.

Errington, A. C., Di Giovani, G. and Crunelli, V. (2014). Extrasynaptic GABAA

Receptors. Heidelberg: Springer.Gassmann, M. and Bettler, B. (2012). Regulation of neuronal GABAB receptorfunctions by subunit composition. Nat. Rev. Neurosci. 13, 380-394.

Hansen, A. J. (1985). Effect of anoxia on ion distribution in the brain. Physiol. Rev.65, 101-148.

Hochachka, P. W. (1986). Defense strategies against hypoxia and hypothermia.Science 231, 234-241.

Hochachka, P.W., Buck, L. T., Doll, C. J. and Land, S. C. (1996). Unifying theory ofhypoxia tolerance: molecular/metabolic defense and rescue mechanisms forsurviving oxygen lack. Proc. Natl. Acad. Sci. 93, 9493-9498.

Hogg, D. W., Pamenter, M. E., Dukoff, D. J. and Buck, L. T. (2015). Decreases inmitochondrial reactive oxygen species initiate GABAA receptor-mediatedelectrical suppression in anoxia-tolerant turtle neurons. J. Physiol. 593,2311-2326.

Howarth, C., Peppiatt-Wildman, C. M. and Attwell, D. (2010). The energy useassociated with neural computation in the cerebellum. J. Cereb. Blood FlowMetab. 30, 403-414.

Hylland, P. and Nilsson, G. E. (1999). Extracellular levels of amino acidneurotransmitters during anoxia and forced energy deficiency in crucian carpbrain. Brain. Res. 823, 49-58.

Johnston, D. and Brown, T. H. (1984). The synaptic nature of the paroxysmaldepolarizing shift in hippocampal neurons. Ann. Neurol. 16, S65-S71.

Krnjevic, K. (1997). Role of GABA in cerebral cortex. Can. J. Physiol. Pharmacol.75, 439-451.

Kuramoto, N., Wilkins, M. E., Fairfax, B. P., Revilla-sanchez, R., Terunuma, M.,Tamaki, K., Lemata, M., Warren, N., Couve, A., Calver, A., (2007). Phospho-dependent functional modulation of GABA(B) receptors by the metabolic sensorAMP-dependent protein kinase. Neuron 53, 233-247.

Larkum, M. E., Watanabe, S., Lasser-Ross, N., Rhodes, P. and Ross, W. N.(2008). Dendritic properties of turtle pyramidal neurons. J. Neurophysiol. 99,683-694.

Lutz, P. L. (1992). Mechanisms for anoxic survival in the vertebrate brain. Annu.Rev. Physiol. 54, 601-618.

Lutz, P. L. andNilsson, G. E. (1997). Contrasting strategies for anoxic brain survival—glycolysis up or down. J. Exp. Biol. 200, 411-419.

Mann, E. O., Kohl, M. M. and Paulsen, O. (2009). Distinct role of GABAA andGABAB receptors in balancing and terminating persistent cortical activity.J. Neurosci. 29, 7513-7518.

Megıas, M., Emri, Z., Freund, T. F. and Gulyas, A. L. (2001). Total number anddistribution of inhibitory and excitatory synapses on hippocampal CA1 pyramidalcells. Neuroscience 102, 527-540.

Melendez-Ferro, M., Prez-Costas, E. and Roberts, R. C. (2009). A new use forlong-term frozen brain tissue: golgi impregnation. J. Neurosci. Methods 176,72-77.

Mody, I. (2001). Distinguishing between GABA(A) receptors responsible for tonicand phasic conductances. Neurochem. Res. 26, 907-913.

Muller, T. (2012). What is the thalamus in zebrafish? Front. Neurosci. 6, 64.Nakane, R. and Oka, Y. (2010). Excitatory action of GABA in the terminal nerve

gonadotropin-releasing hormone neurons. J. Neurophysiol. 103, 1375-1384.Nilsson, G. E. and Lutz, P. (2004). Anoxia tolerant brains. J. Cereb. Blood Flow

Metab. 24, 475-486.Olude, M. A., Mustapha, O. A., Aderounmu, O. A., Olopade, J. O. and Ihunw,

A. O. (2015). Astrocyte morphology, heterogeneity, and density in the developingAfrican giant rat (Cricetomys gambianus). Front. Neuroanat. 9, 1-10.

Pamenter, M. E., Hogg, D. W., Ormond, J., Shin, D. S., Woodin, M. A. and Buck,L. T. (2011). Endogenous GABAA and GABAB receptor-mediated electricalsuppression is critical to neuronal anoxia tolerance. Proc. Natl. Acad. Sci. 108,11274-11279.

Peter, R. E. and Gill, V. E. (1975). A stereotaxic atlas and technique for forebrainnuclei of the goldfish, Carassius auratus. J. Comp. Neurol. 159, 69-101.

Sattler, R. and Tymianski, M. (2000). Molecular mechanisms of calcium-dependent excitotoxicity. J.Mol. Med. 78, 3-13.

Shepherd, G. M. (2011). The microcircuit concept applied to cortical evolution: fromthree-layer to six-layer cortex. Front. Neuroanat.

Shin, D. S. and Buck, L. T. (2003). Effect of anoxia and pharmacological anoxia onwhole-cell NMDA receptor currents in cortical neurons from the western paintedturtle. Physiol. Biochem. Zool. 76, 41-51.

Stensløkken, K.-O., Ellefsen, S., Stecyk, J. A. W., Dahl, M. B., Nilsson, G. E. andVaage, J. (2008). Differential regulation of AMP-activated kinase and AKT kinasein response to oxygen availability in crucian carp (Carassius carassius).Am. J. Physiol. Regul. Integr. Comp. Physiol. 295, 1803-1814.

Tao, W., Higgs, M. H., Spain, W. J. and Ransom, C. B. (2013). PostsynapticGABAB receptors enhance extrasynaptic GABAA receptor function in dentategyrus granule cells. J. Neurosci. 33, 3738-3743.

Ueno, S., Bracamontes, J., Zorumski, C., Weiss, D. S. and Steinbach, J. H.(1997). Bicuculline and gabazine are allosteric inhibitors of channel opening of theGABAA receptor. J. Neurosci. 17, 625-634.

Watanabe, M., Wake, H., Moorhouse, A. J. and Nabekura, J. (2009). Clustering ofneuronal K+-Cl– cotransportes in lipid rafts by tyrosine phosphorylation. J. Biol.Chem. 284, 27980-27988.

Wilkie, M. P., Pamenter, M. E., Alkabie, S., Carapic, D., Shin, D. S. and Buck,L. T. (2008). Evidence of anoxia-induced channel arrest in the brain of the goldfish(Carassius aratus). Comp. Biochem. Physiol. C. Toxicol. Pharmacol. 148,355-362.

704

RESEARCH ARTICLE Journal of Experimental Biology (2017) 220, 695-704 doi:10.1242/jeb.146605

Journal

ofEx

perim

entalB

iology